dsge.xlsx (13.5 KB)
model10.mod (14.7 KB)
Hi, I am trying to estimate a model using bayesian methods, but even though I can generate the IRFs and the equations are consistent with the steady state I have provided in the .mod file, when I proceed to the estimation the following error happens:
The steadystate file did not compute the steady state
Could you please help me spot my mistake?
P.S.: the data in the dsge.xlsx file was (pseudo-)randomly generated and does not correspond to real data. I just want to adjust the dynare-estimation script before preparing the real data.
Your steady state values seem hardcoded. Why should these values hold for all feasible parameter draws during estimation?
I see the problem now, thank you for your (very rapid!) answer, professor Pfeifer.
Another question for the same model, but regarding the stochastic simulations.
The following command (which is working perfectly) will, as far as I can tell, provide the path for the endogenous variables given pseudo-random realizations of the exgenous shocks of the model. Is it possible to exchange the pseudo-random draws by dynare for observed data so I can have a stochastic simulation for the endogenous variables that I can compare with observed data?
stoch_simul(ar=1, order=1, irf=0, periods = 300) Y mu;
Thank you in advance.
Observed data is not the same as observed shocks. Typically you would need to run a filter. See the
Yes, my choice of expression was poor. I meant a series of estimated shocks (from a SVAR, for instance, of other sources, it does not matter). I want a series of shocks for a given exogenous variable and see how the system evolves.